What it is
DNQ–novolak resists are positive photoresists built on a novolak (phenol-formaldehyde) resin loaded with a diazonaphthoquinone (DNQ) dissolution inhibitor; exposure converts the DNQ where light lands into a species that dissolves readily in developer, so the exposed pattern washes away and the unexposed resist remains. The AZ family, Microposit/Kayaku S1800 series, Allresist AR-P and SPR-series resists are all built on this chemistry.
How it patterns
- Substrate
- Resist
- Exposed
The DNQ dissolution inhibitor
Unexposed, the novolak resin alone would already dissolve at a modest rate in an aqueous developer; the DNQ compound dispersed through it acts as a dissolution inhibitor, suppressing that rate by roughly two orders of magnitude so the unexposed film survives development largely intact. The resist's contrast comes almost entirely from removing that inhibition selectively where the mask (or, on a maskless tool, the exposure pattern) lets light through.
Exposure: a Wolff rearrangement, not amplification
At 365 nm or 405 nm, the DNQ absorbs a photon and loses nitrogen in a Wolff rearrangement, forming a ketene intermediate that reacts with trace water in the film to yield an indenecarboxylic acid. That acid is itself soluble in the aqueous-base developer and, more importantly, removes the parent DNQ's inhibiting effect on the novolak around it — so the exposed region's dissolution rate rises sharply. This is a direct photochemical conversion consumed stoichiometrically by the light dose, rather than a catalytic photoacid amplification, which is why DNQ-novolak resists need no post-exposure-bake deprotection step to develop and are relatively insensitive to post-exposure delay.
Develop: the positive tone
Development in an aqueous base — metal-ion-free TMAH, or a KOH/NaOH-buffered developer such as AZ 400K or AZ 340, which thick grades often specify — dissolves the exposed, deinhibited region far faster than the still-inhibited unexposed film, so the pattern that survives is the unexposed pattern: the definition of positive tone. Because the underlying chemistry is a direct photoreaction rather than a diffusing catalyst, the exposed/unexposed contrast is set mostly by the optical dose profile, which is part of why these resists have long been the default for micron-scale mask-aligner and stepper work.
Standing waves and the swing curve
Because the substrate reflects the exposing light back into the film, the standing wave formed between the incident and reflected beams modulates the dose with depth, printing faint horizontal ridges on the sidewall and making the resist's effective sensitivity oscillate with coating thickness (the swing curve) — a datasheet's recommended thickness values sit at a swing-curve extremum for that reason. The swing curve itself is managed by coating at such an extremum or by adding a bottom anti-reflective coating (BARC); a post-exposure bake diffuses the unreacted DNQ enough to smooth the sidewall ridges, but does not remove the swing curve; some AZ-family products are also formulated for a controlled partial image reversal, a separate process built on the same DNQ chemistry that this page does not re-derive (see the image-reversal resist page).
What each step does
- 01 Spin coat
Most DNQ-novolak films are thin and come from a fairly fluid resist, spun on a quick ramp so the film thins while it is still wet and levels flat. The solvent evaporates fast; resist left sitting on the wafer before the spin, or ramped up slowly, can skin over and leave striations.
Adjusting it: Final speed sets thickness, read off the spin curve. Because of the swing curve, a small change in thickness moves the dose needed to clear. Keeping the dispense, ramp and spin time the same on every run matters as much as the target number.
- 02 Soft bake
Drives out most of the casting solvent. Leftover solvent makes even unexposed novolak dissolve faster, so the soft bake directly sets how much the unexposed resist erodes in the developer.
Adjusting it: Under-baked film loses thickness in the developer and adheres poorly. Over-baked film starts losing DNQ to heat before it is exposed, so it needs more dose and develops more slowly.
- 03 Rehydrate
The exposure chemistry needs water: the intermediate formed by light reacts with water in the film to give the soluble acid. The bake drives that water out, and a thick film takes it back from room air only slowly.
Adjusting it: Wait between soft bake and exposure on thick coats. The wait grows with thickness and is longer in dry air. Skip it and the exposed resist develops slowly and unevenly, as if under-dosed.
- 04 Expose
Each absorbed photon converts one DNQ molecule into an indene carboxylic acid and releases nitrogen. There is no catalysis, so the change follows the dose directly. The DNQ also bleaches as it converts, which lets light reach deeper as the exposure goes on.
Adjusting it: Dose sets how much inhibitor is removed at the bottom of the film. Too little leaves resist in the openings; too much clears past the edges and widens them. In thick films the released nitrogen can foam the resist, so the dose is split into pulses with pauses to let it escape.
- 05 Post-exposure bake
Optional, since the chemistry doesn't need it. A short bake lets the unreacted DNQ diffuse a short distance, which smooths the standing-wave ripples on the sidewall.
Adjusting it: Too hot, and heat begins to break down the DNQ in the unexposed resist as well, which lowers the contrast.
- 06 Develop
Aqueous base (TMAH, or a buffered KOH/NaOH developer) dissolves the exposed resist quickly and the still-inhibited unexposed resist slowly. The contrast comes from the ratio of those two rates.
Adjusting it: A stronger developer or a longer develop clears faster but erodes more of the unexposed resist. A weaker developer gives steeper walls but needs more dose. Residue in an opening after a normal develop usually points back at the dose or the rehydration, not at the develop time.
- 07 Hard bake
Optional. Removes the remaining solvent and toughens the novolak so it lasts through wet etching or plating.
Adjusting it: Above the resist's softening point the patterns reflow and their edges round off, which is what you want for rounded profiles and not for vertical ones. A hard-baked film is also harder to strip.
When it goes wrong
Why is the coated film rough, lumpy or silky right after spin coating?
Cause: The resist has aged, was stored too warm, or was thinned too far or with the wrong solvent, so particles have formed in it. Badly degraded resist gels and gives a glossy but bumpy film.
Fix: Replace expired or mis-stored resist, store it as the datasheet says, and dilute only with the manufacturer's thinner, inside its recommended ratio.
Sources: MicroChemicals, Lithography trouble-shooter, Rough or bumpy resist surface after coating; Allresist, FAQs concerning photoresists, Q3, age-related changes
What causes comet-shaped streaks running out from the centre of the coat?
Cause: Air bubbles in the dispensed resist, more rarely particles, are dragged outward by the spin and leave a streak behind them.
Fix: Let the bottle stand after opening or moving it so bubbles can rise out before you dispense, and don't shake it before use.
Source: MicroChemicals, Lithography trouble-shooter, Spin coating: comet-shaped structures
How do I stop the edge bead from causing trouble?
Cause: Surface tension piles resist up at the rim, worse with viscous resists and square substrates. The thick rim dries slowly, spoils contact with a mask and sheds particles.
Fix: Finish viscous resists at a slightly higher speed to thin the bead, or dispense an edge-bead-removal solvent onto the spinning edge. Don't use a fast-evaporating solvent such as acetone for it.
Source: MicroChemicals, Lithography trouble-shooter, Spin coating: edge bead
Why do bubbles appear in the film during the soft bake?
Cause: Put straight onto a hot plate, a thick film releases solvent faster than it can escape, and outgassing substrates add vapour of their own.
Fix: Ramp up to the soft-bake temperature gradually, with a proximity bake or staged hotplates, and pre-bake polymer or other outgassing substrates before coating.
Source: MicroChemicals, Lithography trouble-shooter, Bubbles in the resist film after the softbake
Why does the exposed resist develop slowly or not clear?
Cause: DNQ resists need some water in the film to develop at a normal rate. After the soft bake the film has to take water back from the air, and too short a wait or dry cleanroom air leaves it short.
Fix: Wait between soft bake and exposure, longer for thicker films, and keep room humidity in the moderate range the manufacturer recommends rather than letting it run dry.
Source: MicroChemicals, Lithography trouble-shooter, Development: insufficient development rate, sufficient rehydration?
Why do bubbles or cracks appear in the resist after exposure?
Cause: Exposure releases nitrogen from the DNQ. In a thick film, or under intense light, it forms faster than it can diffuse out and builds up as bubbles and cracks.
Fix: Split the exposure into several shorter doses with pauses between them, lower the intensity, bake the film more thoroughly first, or coat thinner. For very thick work, a resist that releases no nitrogen avoids the problem.
Source: MicroChemicals, Lithography trouble-shooter, Bubbles in resist film after exposure
Why does the wafer stick to the mask during contact exposure?
Cause: Solvent left by a short or cool soft bake keeps the film tacky. Particles or bubbles in the film, hard-contact mode and poor adhesion make it worse.
Fix: Bake to the datasheet schedule, clean the mask, use soft contact where the aligner allows it, and improve adhesion.
Source: MicroChemicals, Lithography trouble-shooter, Sticking between resist film and mask
How do I get rid of standing-wave ripples on the sidewalls?
Cause: Light reflected from the substrate interferes with the incoming light and prints a ripple through the film thickness. On reflective or textured substrates, scattered light also exposes areas it shouldn't, which is notching.
Fix: A post-exposure bake diffuses the unreacted DNQ enough to smooth the ripple. On reflective or textured substrates, a bottom anti-reflective coating cuts the reflection behind the notching.
Sources: Merck, AZ ECI 3000 series datasheet, Post expose bake; MicroChemicals, Lithography trouble-shooter, Accidental exposure of positive resists, textured substrates
Why is the unexposed resist losing thickness in the developer?
Cause: Solvent left by a cool or short soft bake speeds up dissolution. A developer mixed stronger than specified erodes unexposed resist faster than it gains on the exposed resist, and aged resist has lost some of its inhibitor.
Fix: Bake to the recommended schedule for the film thickness, dilute the developer to its specified ratio instead of running it strong for speed, pick a lower-erosion developer where the resist allows, and don't use expired resist.
Source: MicroChemicals, Lithography trouble-shooter, Development: dark erosion too high
Why is a thin residue left in openings that should be clear?
Cause: HMDS applied as a liquid, or onto a cold substrate, modifies the resist next to the substrate so it no longer develops through. A developer that doesn't match the resist can also leave a film.
Fix: Prime with HMDS from the vapour phase onto a heated substrate, never in a spinner that also coats resist, and use the developer the resist maker specifies.
Source: MicroChemicals, Lithography trouble-shooter, HMDS applied correctly; compatible developer?
Why do small features lift or peel during development?
Cause: Water still bound to the substrate surface, badly applied HMDS, organic contamination or humid air all weaken adhesion.
Fix: Dehydration-bake the substrate before coating, prime with vapour HMDS on a hot substrate, clean off organics with acetone then IPA, and keep humidity moderate.
Source: MicroChemicals, Lithography trouble-shooter, Inferior resist adhesion
Why do resist features round off during plasma etching?
Cause: The plasma heats the resist past its softening point and it flows, rounding the profile and moving the edge.
Fix: Keep the substrate cooled during the etch so the resist stays below its softening point, or use a more heat-stable resist for long or high-power etches.
Source: MicroChemicals, Lithography trouble-shooter, Dry etching: resist rounding or flowing
Why does the wet etch undercut further under the mask than expected?
Cause: Wet etchants always undercut a little, but a lot of undercut means the etchant is creeping in under a poorly adhering resist edge.
Fix: Improve adhesion with a dehydration bake and correctly applied HMDS, and hard bake the pattern after development, before the etch.
Source: MicroChemicals, Lithography trouble-shooter, Wet etching: underetching
Why does metal plate in under the resist edge?
Cause: The resist swells in the aqueous plating bath and loses adhesion. Residue left after development, or badly applied HMDS, also spoils contact with the seed layer.
Fix: Optimise adhesion with the usual pretreatment, apply HMDS from the vapour phase, and make sure development leaves the seed metal clean.
Source: MicroChemicals, Lithography trouble-shooter, Electroplating: bad adhesion, underplating
Why won't the resist strip after dry etching or a hot bake?
Cause: Heat, deep-UV light and ion bombardment crosslink the resist, and heated far enough it chars and no longer dissolves in solvent removers. Material redeposited by the etch can also shield it.
Fix: Use a stripper meant for crosslinked novolak, remove redeposited material first, and keep bakes and process temperatures below the point where the resist crosslinks or chars.
Sources: MicroChemicals, Lithography trouble-shooter, Photoresist removal does not work (well); Merck, AZ ECI 3000 series datasheet, Stripping
Where it fits
- For micron-scale positive patterning with wide process latitude, DNQ–novolak resists give predictable, well-characterized contrast across a broad exposure-dose window, which is why decades of mask-aligner and stepper processes are built around them.
- For an aqueous-developer process line, DNQ–novolak resists develop in a metal-ion-free TMAH base or a KOH/NaOH-buffered developer such as AZ 400K or AZ 340; thick grades often specify the buffered kind, so check the recipe's developer against your line.
- For same-day iteration without a fresh photoacid-generator batch to manage, DNQ–novolak resists are comparatively insensitive to post-exposure delay, since the photoreaction is direct rather than catalytic.
- For deep-sub-micron critical-dimension control, DNQ–novolak resists are a poor fit — a direct, non-catalytic photoreaction gives less resolution latitude than that regime demands.
- For very thick single-coat films, DNQ–novolak resists reach a few tens of micrometres per coat — ma-P 1275 to 40 µm, SPR 220-7 to 52 µm; beyond about 50 µm an epoxy or chemically amplified resist is the usual choice.
- For a process where standing-wave sidewall ridges cannot be tolerated, the swing curve has to be managed deliberately — by coating at a thickness on a swing extremum or adding a BARC, with a post-exposure bake to smooth the remaining ridges — rather than assumed away.
Datasheet-cited recipes
Every value on these pages is cited to the manufacturer’s datasheet — spin curve, bake schedule, exposure dose and developer.
Further reading
- F. H. Dill, W. P. Hornberger, P. S. Hauge, J. M. Shaw. Characterization of positive photoresist. IEEE Transactions on Electron Devices (1975). doi:10.1109/t-ed.1975.18159
- J. Pacansky, J. R. Lyerla. Photochemical Decomposition Mechanisms for AZ-Type Photoresists. IBM Journal of Research and Development (1979). doi:10.1147/rd.231.0042
- E. J. Walker. Reduction of photoresist standing-wave effects by post-exposure bake. IEEE Transactions on Electron Devices (1975). doi:10.1109/t-ed.1975.18162
Common questions
What makes a DNQ-novolak resist positive-tone?
Exposure converts the DNQ dissolution inhibitor into a developer-soluble acid, so the exposed region dissolves away in the developer and the unexposed resist remains — the pattern that survives is the unexposed pattern, the definition of a positive tone.
What is the DNQ dissolution inhibitor?
DNQ (diazonaphthoquinone) is a photoactive compound dispersed through the novolak resin that suppresses its dissolution rate in aqueous developer by roughly two orders of magnitude until exposure removes that inhibition; the resist's contrast comes from switching that inhibition off selectively where light lands.
Why do standing waves appear in DNQ-novolak exposures?
The substrate reflects the exposing light back into the film, and the standing wave formed between the incident and reflected beams modulates the dose with depth, printing faint sidewall ridges and making the resist's effective sensitivity oscillate with coating thickness (the swing curve) — which is why datasheets specify a thickness at a swing-curve extremum, or pair the resist with a BARC. A post-exposure bake smooths the sidewall ridges but does not remove the swing curve.
Expose it at 365 and 405 nm
NANYTE BEAM is a desktop maskless lithography system with software-selectable dual-wavelength exposure and 16-bit grayscale — no photomask, no mask cost, same-day iteration.
General photolithography reference material, not a specification of any particular NANYTE BEAM configuration. Product names and trademarks belong to their respective owners; NANYTE is not affiliated with the manufacturers mentioned.